tokenizer_config.json
| 1 | { |
| 2 | "bos_token": "<|startoftext|>", |
| 3 | "clean_up_tokenization_spaces": true, |
| 4 | "eos_token": "<|endoftext|>", |
| 5 | "model_max_length": 1048576, |
| 6 | "pad_token": "<pad>", |
| 7 | "tokenizer_class": "PreTrainedTokenizerFast" |
| 8 | } |